Measuring Research Convergence in Interdisciplinary Teams Using Large Language Models and Graph Analytics
Wenwen Li, Yuanyuan Tian, Sizhe Wang, Amber Wutich, Paul Westerhoff, Sarah Porter, Anais Roque, Jobayer Hossain, Patrick Thomson, Rhett Larson, Michael Hanemann · 2026
Read Measuring Research Convergence in Interdisciplinary Teams Using Large Language Models and Graph Analytics by Wenwen Li, Yuanyuan Tian, Sizhe Wang, Amber Wutich, Paul Westerhoff, Sarah Porter, Anais Roque, Jobayer Hossain, Patrick Thomson, Rhett Larson, Michael Hanemann (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-20204.txt.
Introduction
Li and co-authors track how mixed research teams converge on shared knowledge, using language models and graph tools. Their framework extracts viewpoints from team presentations with large language models. Viewpoints are coded by a Needs Approach Benefits Competition scheme. Inferred flows between presenters form a common semantic base for three analyses. One sorts viewpoints into popular and unique types. Another uses network centrality to measure cross domain influence. A third follows viewpoint movement over time. Expert surveys and cross layer checks guard against model inference errors.
Catalog record
| CREATOR | Wenwen Li, Yuanyuan Tian, Sizhe Wang, Amber Wutich, Paul Westerhoff, Sarah Porter, Anais Roque, Jobayer Hossain, Patrick Thomson, Rhett Larson, Michael Hanemann |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57284 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.20204
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Measuring Research Convergence in Interdisciplinary Teams Using Large Language Models and Graph Analytics Wenwen Li, Yuanyuan Tian, Sizhe Wang, Amber Wutich, Paul Westerhoff, Sarah Porter, Anais Roque, Jobayer Hossain, Patrick Thomson, Rhett Larson, Michael Hanemann Abstract 1 School of Geographical Sciences and Urban Planning, Arizona State University, Tempe, AZ 2 School of Human Evolution and Social Change, Arizona State University, Tempe, AZ 3 School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ 4 Kyl Center for Water Policy, Arizona State University, Tempe, AZ 5 Nicholas School of the Environment, Duke University, Durham, NC 6 Sandra Day O'Connor College of Law, Arizona State University, Tempe, AZ 7 Center for Environmental Economics and Sustainability Policy, School of Sustainability, Arizona State University, Tempe, AZ Correspondence: [email protected] Abstract : Understanding how interdisciplinary research teams converge on shared knowledge is a persistent challenge. This paper presents a novel, multi-layer, AI-driven analytical framework for mapping research convergence in interdisciplinary teams. The framework integrates large language models (LLMs), graph-based visualization and analytics, and human-in-the-loop evaluation to examine how research viewpoints are shared, influenced, and integrated over time. LLMs are used to extract structured viewpoints aligned with the Needs-Approach-Benefits-Competition (NABC) framework and to infer potential viewpoint flows across presenters, forming a common semantic foundation for three complementary analyses: (1) similarity-based qualitative analysis to identify two key types of viewpoints, popular and unique, for building convergence, (2) quantitative cross-domain influence analysis using network centrality measures, and (3) temporal viewpoint flow analysis to capture convergence dynamics. To address uncertainty in LLM-based inference, the framework incorporates expert validation through structured surveys and cross-layer consistency checks. A case study on water insecurity in underserved communities as part of the Arizona Water Innovation Initiatives demonstrates increasing viewpoint convergence and domain-specific influence patterns, illustrating the value of the proposed AI-enabled approach for research convergence analysis. Keywords : Artificial Intelligence (AI), Large language models (LLMs), convergence research, Human-in-the-loop, water insecurity, quantitative-qualitative 1. Introduction In recent years, the interdisciplinary nature of scientific research has advanced into a convergence research paradigm (Petersen et al. 2023). Convergence research emerges as an expanded form of interdisciplinary research (Finn et al. 2022), integrating knowledge, insights, methodologies, tools, and mindsets from diverse disciplines through collaboration. This research paradigm can produce innovative solutions for complex real-world problems that cannot be solved when considered through a single disciplinary lens, such as disaster response and water insecurity challenges (Peek et al. 2020; Wutich et al. 2023). Convergence research has become a transformative framework in scientific research, advocating for deep collaboration and integrated solutions to sustain scientific innovation (Baaden et al. 2024). It also serves as a fundamental mechanism for bridging the gap between theoretical research and practical applications (Bukvic et al. 2022; Lobo et al. 2021). Additionally, convergence research fosters new knowledge and innovation, driving scientific and technological advancements. To support convergence research and scientific innovation, government agencies such as the US National Science Foundation (NSF) and European Union (EU) have taken initiatives on new programs. The EU’s Horizon Europe programme has a convergence strategy, where the research focuses on tackling climate change; NSF’s Growing Convergence Research and Convergence Accelerators aim to incubate convergence across scientific disciplines, translating basic research into solutions with real-world applications, thus opening new frontiers in science and engineering. The convergence process is also about the researchers themselves, as the process of working closely with scholars from other disciplines broadens their thinking and should provide benefits to future research. Therefore, besides seeking advances in science and societal impacts, these efforts also call for innovative approaches for measuring interdisciplinary team’s research convergence. Despite the acknowledged benefits of convergence research, there are significant challenges in measuring the effectiveness and depth of convergence within research teams (Alwin and Hofer 2008; Feller 2006; Petersen et al. 2021). Most related studies have typically focused on measuring interdisciplinarity and collaboration based on scholarly output, such as scientific publications (Rinia et al. 2002; Rinia 2007; Raan 2005). However, due to the significant time lag in publications, this approach has limitations in capturing the micro-dynamics of the convergence process, hindering the ability to conduct systematic, near real-time reflection on the team's progress as the collaboration continues. Moreover, several empirical analyses have shown that other forms of data, such as questionnaires assessing team collaboration, can provide insights into the convergence process (Hubbs et al. 2020). This suggests that convergence detection can be more timely and should incorporate more diverse data sources rather than solely relying on publications, an area that deserves further investigation. To address the challenges of detecting and monitoring an interdisciplinary team’s research convergence, our research addresses a significant gap in the existing literature by pioneering an Artificial Intelligence (AI)-driven methodological framework that captures the dynamic and real-time interactions among convergence participants from various domains. We adopted and extended a value creation framework, termed NABC (Needs, Approaches, Benefits, and Competition; Carlson and Wilmot 2006) to practice and record the team’s convergent thinking over time. Originating in the research and development sector, this framework has been effectively adapted for problem-solving in other areas, such as in addressing complex environmental challenges (Westerhoff et al. 2021). By structuring presentations and discussions around the NABC components, research teams are compelled to articulate and align on the specific societal needs they aim to address, the innovative approaches they will employ, the benefits relative to existing solutions, and the competitive landscape. NABC also provides a common structure, without any particular disciplinary bias, to aid the convergence process. Specifically, we developed a multi-layer, AI-driven analytical framework to examine research convergence in interdisciplinary teams. This study analyzes team convergence and collaboration from complementary perspectives and addresses three research questions: (1) What viewpoints are widely shared across the interdisciplinary team, and which remain unique or domain-specific? (2) How do research viewpoints from different domains influence one another within the interdisciplinary team? (3) How do team knowledge and viewpoints progressively integrate and converge over time? The proposed qualitative-quantitative framework integrates large language models (LLMs) with graph-based analytics and visualization. LLMs are used to extract structured viewpoints and infer potential opinion flows through carefully designed and constrained prompts, while graph-based methods support similarity analysis, cross-domain influence assessment, and temporal convergence analysis. To ensure reliability, expert-level evaluation is incorporated through human-in-the-loop validation, and analytical results are further evaluated through cross-layer consistency among multiple analytical perspectives. To the best of our knowledge, this work represents one of the first efforts to provide an automated qualitative-quantitative framework for analyzing interdisciplinary research convergence that captures both structural relationships and longitudinal dynamics. The remainder of this paper is organized as follows. Section 2 reviews related literature on research convergence analysis. Section 3 introduces the interdisciplinary team and the convergence research problem of water insecurity in underserved communities. Section 4 presents the methodological framework integrating large language models, graph-based analytics, and human evaluation. Section 5 reports and analyzes the results. Section 6 discusses results evaluation and the mitigation of LLM-related uncertainty. Finally, Section 7 concludes the paper and outlines directions for future research. 2. Literature review 2.1 Measuring research convergence through bibliometric analysis Many studies have attempted to measure interdisciplinarity and collaboration, and the common approach is bibliometric analysis (Leydesdorff and Ivanova 2021; Shi and Wang 2022). There are two